Human chromosomes end in telomeres, 5–15 kilobases of double-stranded (TTAGGG)n repeats that culminate in a 3′ single-stranded overhang [1,2]. Facilitated by the six-subunit protein complex shelterin, this overhang folds back and invades the proximal telomere, creating a telomere loop (T-loop) that protects the linear end of the chromosome from recognition by the DNA damage response machinery [3,4]. Shelterin consists of telomere repeat-binding factors 1 and 2 (TRF1/TRF2), repressor–activator protein 1 (Rap1), TRF1-interacting nuclear protein 2 (TIN2), protection of telomeres 1 (POT1), and TIN2 and POT1-organizing protein (TPP1). TRF1 and TRF2 bind double-stranded DNA and are highly specific for the telomere repeat sequence. In contrast, the POT1 protein binds single-stranded DNA in the 3' overhang and the local displacement loop formed during formation of the T-loop [3]. Incomplete DNA replication of telomere ends leads to erosion of the telomeric DNA following each cell division. Telomere erosion keeps the proliferative capacity of cells in check, as critically short telomeres become deprotected and susceptible to erroneous non-homologous end joining that can drive cells into crisis [5]. Although the majority of cancers require activation of a telomere maintenance mechanism to overcome this limitation on proliferation, a subset of cancers have no detectable telomere maintenance mechanism [6]. This phenotype is known as ever-shorter telomeres (EST). In some cases, these tumors begin with longer than average telomeres, indicating a possible lengthening event prior to transformation. However, the mechanism of telomere elongation is not maintained, and these cells continue to proliferate even as telomeres undergo continuous shortening [7,8].

Apart from those displaying the EST phenotype, the majority of human tumors promote telomere elongation through reactivation of the enzyme telomerase or the alternative lengthening of telomeres (ALT) pathway. Unlike the reverse transcriptase activities of telomerase, ALT relies on homologous recombination to promote telomere elongation [9]. The prevalence of ALT across all cancers is estimated to be only around 4%; however, ALT rates in some cancers of neuroepithelial and mesenchymal origin are likely to be over 50% [10,11]. These tumors include sarcomas, gliomas, and pancreatic neuroendocrine tumors that have few effective treatment options and overall poor prognoses, highlighting a need for the development of novel strategies that target the ALT pathway.

Mechanistically, ALT relies on a type of recombination referred to as break induced replication (BIR) to promote telomere elongation. BIR is a pathway that repairs one-ended double-strand breaks (DSBs), which arise when challenges to the replisome cause replication fork stalling. These telomeric DNA DSBs are clustered into nuclear bodies that contain the scaffolding protein promyelocytic leukemia (PML) protein [12]. These ALT-associated PML bodies (APB) provide the framework for the recruitment of the DNA replication and recombination proteins required to initiate BIR [13]. The recombination intermediates generated at ALT telomeres can be resolved in a way that leads to crossover events in the form of telomere sister chromatid exchange (T-SCE) [14]. As a byproduct of the ongoing recombination, ALT tumors generate extra-chromosomal, partially single-stranded, C-rich, circular DNA products known as C-circles. C-circles are exclusive to ALT-positive tumors and have become a reliable marker of ALT status [15].

It is unclear exactly how ALT is activated in cancer; however, the accumulation of replication stress within telomeric DNA has been implicated in the process. Telomeres are common fragile sites of the genome and sources of endogenous DNA replication stress. In addition to polymerase slippage caused by the repetitive sequence, the T-loop and G-quadruplexes that form in the G-rich strand can act as physical barriers to the replisome and lead to stalling of the replication fork [16]. Additionally, telomeres are origin-poor, with the majority of replication initiation events likely originating in the subtelomeric region [17]. Compared with other regions of the genome, stalled telomeric forks are therefore less likely to be rescued by a converging fork or dormant origin firing. Irreparably stalled forks can then lead to double-strand DNA breaks that are repaired via BIR.

The replication stress at ALT telomeres is likely aggravated by changes to the chromatin landscape. The most common ALT-associated mutations occur in the chromatin remodeling enzyme ATRX (α-thalassemia/mental retardation syndrome X-linked) and its histone chaperone partner DAXX (death domain associated protein), which form a complex to deposit H3.3 in regions of heterochromatin including telomeric DNA [18]. Defects in ATRX and/or DAXX function lead to reduced H3.3 deposition, defects in heterochromatin formation, and often the accumulation of non-canonical, or variant, repeats at ALT telomeres [19, 20, 21, 22, 23]. In addition, loss of ATRX also leads to an increase in the long non-coding RNA transcribed from telomere ends, TERRA [24]. The increase in TERRA in ALT-positive cells leads to the formation of RNA:DNA hybrids, or R-loops, that can drive collision between the transcription machinery and the DNA replication machinery, exacerbating replication stress [25]. Inhibiting TERRA metabolism has been shown to decrease ALT phenotypes, indicating that formation of these R-loops may contribute directly to the regulation of ALT-mediated DNA synthesis [26].

Collectively, defining the ALT mechanism has allowed researchers to begin to identify critical steps within the pathway, highlighting vulnerabilities that could be exploited therapeutically. The emerging trend is that cancer cells rely on a certain balance of ALT activity, where inhibition or exacerbation of this activity can compromise cell viability. Here, we will discuss recently identified vulnerabilities in ALT cancer cells and highlight how these vulnerabilities may provide the foundation for the development of ALT-targeted therapies.